If you saw our earlier post where we performed an analysis on ROA (Return on Ada) for pools vs their active stake, you would have seen where we visualized what we called the ‘small pool tax’. This refers to the negative effect on returns that the 340 fixed fee per epoch disproportionately has on smaller pools vs larger ones.
Today we are implementing a fee change based on this analysis.
Effective immediately we are reducing our fee to 1% and will implement scheduled increases based on our achieved stake. See below for a more detailed explanation as to how this will work.
If you look at the chart below, it shows the theoretical ROA based on pool size, taking into account this 340 ADA fee.
As you can see, even at different fee structures, all small pools suffer from this penalty introduced by the 340 ADA fee, however the effect of this fee is greatly reduced as the pool gets larger, eventually flattening out and having a negligible effect as the pool approaches saturation.
In order to be fair to our delegators, and not penalize their ROA while we are still relatively small, we are implementing the schedule we proposed in an attempt to maintain a minimum (theoretical) ROA for our pool. We specify theoretical, because luck is till a large factor while d is high, and the value of individual blocks is still high.
Below is our new fee schedule, implemented in epoch 216 (taking effect epoch 218)
|Pool size < 13M ADA||1%|
|Pool size – 13M < 19M||2%|
|Pool size – 19M < 23M||3%|
|pool size – 23M +||3.5%|
Using this structure, we can significantly offset the tax imposed by the minimum fee and provide our delegators a much improved ROI while we are still relatively small. The theoretical ROA given the above schedule is illustrated below:
As you can see, once we reach 10M we will reach our minimum desired theoretical ROA, and we maintain this minimum until we reach 23M at which point our ROA will continue to trend upwards.
Let us know what you think of this change, we’ve already implemented this prior to the epoch change tomorrow.